Nirav Patel

Training Deep Multimodal Neural Networks 🧠

Hi! I'm a final year Master's student at L. D. College of Engineering, where I'm advised by Prof. Payal Prajapati. My research focuses on developing computer vision models that can understand and learn long videos. To this end, my recent work explores learning multimodal representations from videos and textual queries to summarize long videos influenced by textual queries.

Previously, I completed my undergraduate studies in Computer Science and Engineering from Government Engineering College, Patan.

Email  |  Github  |  Google Scholar  |  Blog

profile photo

News

[Dec 2024] Presented my first research work and chaired Image/Video Analysis session at ICARCV 2024, Dubai, UAE.

Research

Your Interest, Your Summaries: Query-Focused Long Video Summarization

Nirav Patel, Payal Prajapati, Maitrik Shah
ICARCV 2024

Paper | Website | Code
Query Focused Video Summarization PyTorch Implementation.

Shuwen Xiao, Zhou Zhao, Zijian Zhang, Xiaohui Yan, and Min Yang. "Convolutional Hierarchical Attention Network for Query-Focused Video Summarization." (AAAI 2020)

Paper | Code
Micrograd Autograd Automatic Differentiation Engine implementation.

Blog | Code

Professional Service

  • Session Chair: ICARCV 2024 Image/vision session
  • Teaching Assistant: Programming in C

Open Source Contributions

I have made 7 pull requests that have been merged into the code base of pandas. Contributions mainly include fixing bugs in sum and concat operations, and documentation improvements. Added tests for already fixed bugs, ensuring enhanced code quality and reliability.

Teaching

In my final year of my Master's program, I instructed bachelor's students in C programming and Data Structures and Algorithms at L. D. College of Engineering ❤️. Pic: Taking a lab session, explaining the concept and implementation of linked lists.

Blog

Dec 2023 Neural Networks: Fits Squiggle To The Data
Jun 2023 Sharing my experience of open source contribution and internship on setting up projects locally


Cloned from JanBarron.